AI Paper Atlas

Research map / Adversarial robustness and security / Certified and empirical robustness

Rethinking Fast Adversarial Training: A Splitting Technique To Overcome Catastrophic Overfitting

Masoumeh Zareapoor, Pourya Shamsolmoali

ECCV 2024 · Adversarial robustness and security › Certified and empirical robustness · approach: Adversarial training refinements · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add DR splitting (RAT) to MEP-PGI at eps=16/255MEP-PGI · CIFAR10 (ResNet20, RA-AA)RA-AA (%): 19.31 -> 23.91Improved
add DR splitting (RAT) to MEP-PGI at eps=16/255MEP-PGI · CIFAR10 (ResNet20)standard accuracy (%): 54.23 -> 69.86Improved
add DR splitting (RAT) to Fast-AT at eps=16/255Fast-AT · CIFAR10 (ResNet20, RA-AA)RA-AA (%): 14.97 -> 19.64Improved
add DR splitting (RAT) to PGI-BP on Tiny-ImageNetPGI-BP · Tiny-ImageNet (PreActResNet18)RA-AA (%): 4.18 -> 6.25Improved
add DR splitting (RAT) to PGI-BP on ImageNetPGI-BP · ImageNet (ResNet50)RA-PGD10 (%): 2.81 -> 8.70Improved
replace multi-step PGD-AT with RAT-MEP (single-step) on CIFAR10RAT-MEP · CIFAR10 (ResNet20)RA-PGD50 (%) and time: 34.21 -> 33.64 RA-PGD50; 6.3 -> 1.7 hoursMixed
replace PGD-AT with RAT-MEP on CIFAR100 AutoAttackRAT-MEP · CIFAR100 (ResNet20)RA-AA (%): 12.54 -> 11.49Hurt
apply RAT to OAATOAAT · CIFAR10 (ResNet20)RA-PGD-10 SQ (%): 21.43 -> 24.66Improved
compare RAT to Fast-GA at small eps=8/255 (WideResNet34-10)RAT · CIFAR10 (eps=8/255)AA (%): 46.37 -> 46.21No clear effect
use higher regularization strength lambda (0.9 or 1) (ablation)RAT · CIFAR10 (PreActResNet18, validation 1024 images)PGD-20 accuracyMixed
use lower lambda with large eps (ablation)RAT · CIFAR10 (PreActResNet18)PGD-20 accuracyHurt
vary lambda within 0.3 to 0.6 (ablation)RAT · CIFAR10 (PreActResNet18)PGD-20 accuracyNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy